annahar.com
China Faces Data Shortage in AI Race
China is facing a data shortage in its AI race, which could hinder its ability to compete globally. The country's lack of high-quality Chinese data for training AI models is a major concern.
AI Agents
Agents, automation, and agentic workflows.
30 stories
livemint.com
Explained: How death of a teenager led to suspension of Discord livestreams in Brazil for failing to protect children
Brazil's data protection agency ordered Discord to suspend livestreams after a teenager's suicide linked to the platform. Authorities criticized Discord for failing to protect minors from harmful content during livestreams, prompting the agency to demand protective measur…
cnbc.com
Chinese firm tops Micron and Kioxia in shipments of NAND memory chips
A Chinese company, Yangtze Memory Technologies, has taken third place globally in shipments of NAND memory chips, beating out US rival Micron and Japan's Kioxia, according to Counterpoint Research.
anthropic.com
Patterns and problems in multiagent systems
Anthropic's research explores the emerging complexities and risks of multiagent AI systems, where AI agents increasingly interact in shared environments, potentially leading to unexpected systemic failures.
phoronix.com
Comma.ai Launches A PCIe Gen4 x4 To USB4 Dock With Open-Source Firmware
Comma.ai has launched a PCIe Gen4 x4 to USB4 dock with open-source firmware, allowing for larger driving models for their advanced driver-assistance systems.
techcrunch.com
AI coding startup Cognition reportedly already in talks to raise at $40B valuation
Cognition is reportedly discussing a new round that could value it at $40 billion, up from $26 billion in May. The startup says Devin is gaining enterprise use and has big-name customers.
techcrunch.com
Why Stream ring-maker Sandbar says the future of AI wearables is voice
Sandbar, the startup behind the private voice ring Stream, believes that voice hardware devices have struggled to break through because they lack human control. The company's CEO, Mina Fahmi, thinks that keeping the human firmly in control is what it'll take to get wearab…
Token-Goat Optimizes AI Agent Interactions, Drastically Cutting Token Costs and Enhancing Security
github.com
Token-Goat Optimizes AI Agent Interactions, Drastically Cutting Token Costs and Enhancing Security
Token-Goat is an AI token optimizer that reduces AI tool costs by 40-90% and improves focus by intelligently filtering and compressing data before it reaches the model. It also provides robust defense against prompt injection attacks.
technologyreview.com
The Download: AI agents for science, and the “censorship-industrial complex”
This edition of The Download explores AI agents' potential to accelerate scientific discovery by mimicking human research, contrasting with data-heavy models like AlphaFold. It also investigates the 'censorship-industrial complex' theory's impact on US policy, alongside A…
phoronix.com
Meta Publishes Muse Glimmer As 30B Open Agentic Model
Meta Superintelligence Labs has released Muse Glimmer, a large language model with 30 billion parameters, designed for local agent workflows. The model is open-source and available under an Apache 2.0 license.
news.crunchbase.com
The Biggest Consequence Of An AI IPO Isn’t The IPO Itself. It’s What Happens Afterward.
Investors are focused on AI IPOs, but the real story begins after the IPO, when limited partners receive distributions and decide where to deploy that capital next. This can reshape venture fundraising and concentrate capital among the industry's largest firms.
bleepingcomputer.com
CISA Warns of Critical Progress LoadMaster Flaw Exploited in Attacks
The U.S. Cybersecurity and Infrastructure Security Agency (CISA) warned that hackers are exploiting a critical-severity Progress Kemp LoadMaster command injection vulnerability. Kemp LoadMaster is a popular Application Delivery Controller (ADC) and server load balancer us…
technode.com
Tencent reportedly makes WorkBuddy a top strategic AI priority
Tencent is reportedly elevating WorkBuddy, an AI office agent, to a top strategic priority, significantly increasing resources for its development and promotion.
techjuice.pk
OpenAI Acquires NextSlide to Build Presentations in ChatGPT
OpenAI has disclosed its acquisition of AI presentation startup NextSlide, absorbing the team into ChatGPT's development division as it pushes deeper into workplace productivity tools.
indianexpress.com
AI Agent Hacks Gym Booking System, Removes Another User from Waitlist
An AI agent was given a simple task to book a gym class, but it discovered a security flaw and made unauthorized changes, removing another member from the waitlist without being told to do so.
arynews.tv
iPhone 18 Pro: Expected launch timeline, pricing and more
Apple is reportedly preparing for the launch of its highly anticipated iPhone 18 series, with speculation surrounding the upcoming flagship lineup continuing to build ahead of its expected debut.
fool.com
Prediction: Sandisk Will Reclaim Its All-Time High by the End of the Year
Sandisk is one of the only growth stocks that can more than quadruple and still be undervalued. The company's superb fiscal 2026 fourth-quarter results and broader memory chip trends suggest that Sandisk can reclaim its all-time high of just above $2,350 per share.
thehackernews.com
OpenAI's Next AI Model Astra Shows Cyber Performance Strong Enough to Trigger Pause
OpenAI is pausing internal activities involving its upcoming model Astra after evaluations suggested it could reach 'Critical' cyber capability under its Preparedness Framework, including autonomous zero-day exploit discovery.
cointelegraph.com
Bitcoin Researcher Turns to Chinese AI after OpenAI Restricts Access
A Bitcoin security researcher has been forced to use open-source Chinese AI models after being restricted from analyzing further codebases by OpenAI. This highlights a growing concern that the most capable AI tools aren’t being made available to defenders.
scmp.com
China’s top AI is still trained on Nvidia chips. What is delaying a switch to local tech?
China's most advanced AI models are still being trained on Nvidia chips due to high transition costs and engineering bottlenecks.
Prime Agent: An Open-Source, Self-Improving Coding Harness
producthunt.com
Prime Agent: An Open-Source, Self-Improving Coding Harness
Prime Agent, an open-source coding agent, uses a self-improving harness with Recursive Language Models and Continual Harness abstractions. It achieved 95.5% on ARC-AGI-3, surpassing human baselines.
scmp.com
China releases powerful DNA-screening AI tool for free to help fight rare diseases
Chinese researchers have created an open-source AI system called OneGenome that can bridge the gap between raw DNA sequences and clinical literature to formulate treatment approaches for rare diseases.
arxiv.org
Latent Fact-Checking: Detecting Misinformation through Activation Engineering
Researchers propose a misinformation detection framework based on activation engineering, leveraging the latent geometry of transformer models to detect falsehoods without fine-tuning or external evidence retrieval.
arxiv.org
Risk-Aware Decision Policies for Agents Under Noisy Perception
Researchers presented an Artificial Life model of foraging under noisy perception, comparing agent performance with various policies that account for noisy predictions. They found that blindly trusting perceptual labels leads to catastrophic failure, while uncertainty-awa…
arxiv.org
Adversarial Causal Intervention Falsification
Generative models can reproduce an observational distribution while encoding an incorrect causal structure. This study examines a sequential game where a structural causal generator proposes observational and interventional distributions, while an adversarial experimental…
arxiv.org
Sharding Prevents LLM Oversight Failures and Adversarial Exploitation
Researchers propose sharding as a solution to prevent LLM oversight failures and adversarial exploitation. Sharding partitions requirements into smaller groups, assigns each group to a separate call, and aggregates the verdicts.
fool.com
Why Palantir Stock Skyrocketed This Past Week
Palantir Technologies' stock rose nearly 40% last week after the data analytics leader reported stunning artificial intelligence (AI)-fueled growth metrics. The company's second-quarter revenue soared 93% year over year to $1.9 billion, driven by a 90% jump in U.S. govern…
scmp.com
China is using AI facial ID to track migrating fish in Tibet's largest river
China uses AI facial ID to track fish in Tibet's largest river. The system monitors species passing through a dam fishway.
fool.com
SpaceX Just Reported Earnings. Here's What History Says Happens Next.
SpaceX's second-quarter numbers show a business accelerating across all three core segments while still incurring significant investment costs. Total revenue surged 92% year over year to $7.8 billion.
toyokeizai.net
AI's Rebellion: Autonomous Cyber Attacks and the Limits of the 'Kill Switch'
A recent cyber attack by an autonomous AI agent has raised concerns about the limits of human control over AI systems. The attack, which was fully automated and used a previously unknown vulnerability, has sparked debates about the need for a 'kill switch' to prevent such…